The 12th biennal collective intentionality conference — july 13–25, 2020
Team reasoning and collective moral obligation
In certain situations, it feels apt to ascribe a collective moral obligation to a group even if the group is unstructured and lacks the properties usually required for moral agency. For example, consider a group of unrelated bystanders in a subway car witnessing an assault (a widely discussed case from Held 1970). Intuitively, it seems that the group has a moral obligation to stop the assault even if no individual interference would be helpful on its own. Most explanations of this intuition have in common that they separate the subject of the obligation – the collective – from the putative addressee of the moral demand, with a capacity for moral deliberation – the individual or the potential group agent. We find this type of separation problematic. Unstructured groups lack a decision procedure, but there is a sense in which they can engage in collective deliberation. We suggest that by addressing the group with a collective demand, we appeal to the individuals’ capacity to regard the situation from a group perspective and to deliberate about what the group should do, to “team reason”. In that sense, the gap between the subject of the obligation and the moral deliberator can be bridged. If this is what we do when we assign an obligation to an unstructured group, is that obligation collective or distributive? Since only the collective has the ability to do what is morally required, the collective is the subject of the obligation. Our claim is that it makes sense to address such a group collectively and demand that it acts as a collective. When we do so we appeal to each individual’s capacity to frame the choice from the group perspective, i.e. to identify with the group and reason as a team member. Under which conditions would it be reasonable to demand that a group of unrelated bystanders choose a course of action requiring them to view the situation from a group perspective? This will depend on what it takes for an individual to identify with a group, whether such a change of perspective can be voluntarily chosen at all, and if the change can be rationally motivated. These are controversial issues, and we briefly explore the moral implications of the main positions in this debate.
Discussion
12 thoughts on “Team reasoning and collective moral obligation”
Leave a Reply
You must be logged in to post a comment.
Only ISOS members (with valid fee) can access to the discussion section. Visit ISOS website to subscribe.
sponsors
Olle and Bjorn, thanks for the talk. Sorry I couldn’t participate in the live session. I am wondering whether team reasoning and group identity are necessary to understand how individuals come to work together to do what they should when they are not already organized for action. I agree that we can say in the case of burning building that the two of them have an obligation to save the two children, or more generally as many as they can. But it seems to me that we can give this a pretty deflationary reading. Each of them has an obligation to do what he can to save as many children as he can. Neither can do it alone but each knows that together they can do it. So neither can get off the obligation to do something on the grounds that he can’t do anything about it. Each can do something about it. Each can contribute to their coming to share an intention to save the two children. So each minimally has an individual obligation to do what he can to make that the case. And since communication is cheap, the first step is just to make a proposal to adopt a plan to do what is appropriate. I don’t think this involves any group identification in any sense other than being willing to do something intentionally with the other to save the children. I don’t see that it really requires them to adopt the group perspective either. The claim that *they* ought to save the two children can be read: Each of them has an obligation to contribute to their saving the children. So the subjects of what obligations there are in this case are individuals. And we derive that distributive obligation from their individual obligations to do what they can together with the fact that in the circumstances the only route to saving the children involves cooperating with the other to do so.
Thanks Kirk for this challenge/comment. You write: ”The claim that *they* ought to save the two children can be read: Each of them has an obligation to contribute to their saving the children.”
We agree that in many situations, we can simply understand “they have an obligation” distributively in the way you suggest, perhaps in line with Regan’s principle that every individual has an obligation to co-operate with whoever else is willing to co-operate, in order to produce the best result. But on the face of it, your proposal requires a reinterpretation when it comes to the type of case we discuss. We want to the make sense of the apparent ascription of a single obligation to an unorganized group. It also seems to us that intuitively, *they* may have an obligation to e.g. save the children even if it isn’t the case that each of them has the obligation to contribute (here we are following e.g. Michael McKinsey and Gunnar Björnsson). Suppose one of Ale or Olivier for some reason doesn’t group-identify, or does group-identify but is unwilling to contribute for whatever reasons (perhaps he doesn’t want his clothes to smell of smoke). Here, the other’s unwillingness will cancel’s the other’s obligation to contribute, but not their collective obligation to save the two children. Or at least, this is how it seems to us.
Björn, thanks for the reply and further explanation. So you’re thinking (1) is true
(1) They (A and O) have an obligation to save the two children.
even if
(2) O does not have an obligation to do anything to save the children
because (he knows that)
(3) A will not help.
But if (1) means (1D) it is not compatible with (2).
(1D) Each of them has an obligation to contribute to their saving the two children.
So (1D) is not the proper analysis of (1), at least as it is to be understood in this case.
Now, I am supposing that the idea that (3) in the circumstances entails (2) rests on the idea that ought implies can, and if O cannot do anything to save the children, O does not have any obligation to do anything to save the children.
For a clear case we have to imagine that there is nothing that O could do that would even increase the probability that they would make an effort to save the children. In many circumstances success of joint endeavors is not assured but this doesn’t relieve us of our obligations. And when the stakes are high as they are in this case, even a small chance is worth striving for.
I think it is hard to think of a natural case in which the conditional epistemic probability that O assigns to saving the children on trying to get A to cooperate is zero, and even harder to think of one in which O should think that the conditional probability was zero.
I am inclined to say that, when we think about cases in which A refuses to help, we are thinking of more or less natural cases. It is hard not to let our thinking about cases be informed by our understanding of human agents. I think we do think in those cases that they have an obligation to save the two children. A’s refusal does not relieve him of his obligation. But I think we tend to think there is not zero chance that O will fail to convince him, and so he still has an obligation.
If O knew A were in a wheelchair, then O would not have an obligation to convince him to help, because he couldn’t. But then I think we would also say that they did not have an obligation to save the children. If A were pathologically afraid of fire and could not enter a burning building, I think we’d say the same thing.
What we need is a case in which we conceive of A as able to help and to contribute, but where he does not do so of his own free will, and there is zero chance that he could be persuaded, and then see what our judgments are. Maybe A set the fire and aims to kill the children? But even in that case it is not clear there might not be considerations that would override his settled plan. Maybe we think that if someone refrains from doing his duty freely, there is always the chance that he might have chosen otherwise on some consideration.
All of this is to say that I can’t quite think myself into a case that has the relevant features to see whether I have the relevant intuition.
Thanks Kirk! We will think more about this and try to reply in greater detail after the holidays. Some preliminarily thoughts. We think that our proposal is more deflationary than you seem to understand it. Also, we are not sure that we need a case where the “conditional epistemic probability that O assigns to saving the children on trying to get A to cooperate is zero [or] one in which O should think that the conditional probability was zero”. Wouldn’t it suffice with a case where there is a discrepancy between what A should do as part of what they should do and what A should do given what A believes, possibly with a low degree of confidence, that O will do, even after attempted persuasion if that possibility is available? Such cases do seem less rare.
Cheers
Olle & Björn
So, just to write down the final thought I was having in the live session. I was wondering what would be involved in adopting the we-frame in these situations. In the general run of cases that are considered in the team-reasoning literature, there are two characteristic steps of “adopting the team perspective”. (i) we are choosing in the first instance between action-types that are whole patterns of actions of individuals (and then doing our part in the chosen pattern) (ii) there is a new set of utilities defined over these pattern types (often modelled by some aggregation over individual utilities).
Now, in your parade case (burning building) you get both characteristics (i) and (ii) of adopting the we-frame, since there’s no disagreement among the participants as to the value of outcomes of the available patterns of actions.
But there will be cases, of course, where the individuals involved disagree over the ranking of patterns of actions. You could imagine some evil participant who wants the kids to die in your original scenario; or construct different situations where moral disagreement is more reasonable. Here’s an assumption I’m making: the fact that there’s no consensus in these cases doesn’t mean the collective obligation goes away. After all, in the individual case, the fact that a murderer wishes his victim to die doesn’t effect the moral imperative that he *not* kill victim. It feels like the same should go of the collective case. If one of a group whom wants the victims to die in the burning house, that shouldn’t alter the fact that there’s a moral imperative on the group to save the victims (all it means it that the imperative is going to be violated).
It is part of your view that it’s a condition on collective obligation that a we-frame is available to the individuals within the collective: that they “have the capacity to identity with the group and view the situation from the group’s perspective”.
My worry is the following: it’s not clear to me that a we-frame will be available to members of a collective who have no moral consensus about how to rank options. The problem is that in conditions of dissensus, there’s nothing which is “the group’s perspective”. To be sure, an individual could contemplate various patterns of action, and evaluate those patterns by deploying *their own* moral judgements. But that’s to apply their own moral perspective, not anything recognizable as the group’s perspective.
We could accept this, and say there’s no we-frame perspective in cases of moral dissensus, but if we hold onto your precondition on what it takes for there to be a moral obligation in this case, then we end up having to deny that the two people were obliged to save the child.
Thanks Robert for elaborating on and writing down your question in more detail! This is very helpful.
You write: “In the general run of cases that are considered in the team-reasoning literature, there are two characteristic steps of “adopting the team perspective”. (i) we are choosing in the first instance between action-types that are whole patterns of actions of individuals (and then doing our part in the chosen pattern) (ii) there is a new set of utilities defined over these pattern types (often modelled by some aggregation over individual utilities).”
That’s right, although one might stress that the new set of utilities resulting from the shift of agential perspective must not necessarily reflect an aggregation of individual utilities, but will depend on the team-reasoner’s we-preferences about the options available to the group.
You write: “My worry is the following: it’s not clear to me that a we-frame will be available to members of a collective who have no moral consensus about how to rank options. The problem is that in conditions of dissensus, there’s nothing which is “the group’s perspective”. To be sure, an individual could contemplate various patterns of action, and evaluate those patterns by deploying *their own* moral judgements. But that’s to apply their own moral perspective, not anything recognizable as the group’s perspective.
We could accept this, and say there’s no we-frame perspective in cases of moral dissensus, but if we hold onto your precondition on what it takes for there to be a moral obligation in this case, then we end up having to deny that the two people were obliged to save the child.”
The shift of agential perspective takes place “in the head of an individual” and we argue that the members’ capacity for taking the group’s perspective is a necessary condition for assigning a collective obligation to the group. But even if all group members have this capacity, the group may fail to fulfil their obligation for various reasons. One such reason could be that group members have immoral preferences and think that “we should let the kids die” even when they regard the situation from the group’s perspective. In that case our account would be compatible with saying that the group had the obligation and should be blamed for not fulfilling it.
In the type of case that you describe, one could perhaps argue that the composition of the group, with salient disagreement about how to evaluate the group’s options, prevents members from identifying with the group in the first place (if we assume that the disagreement is out in the open). Alternatively, we could say that the ”group’s perspective” just refers to the agency transformation and to the consideration of patterns of actions as making up the options among which the members choose, and that the ranking and evaluation of those options which determines the collective obligation are given by the correct moral theory. If we go for the former option, then assigning a *collective* obligation to do the right thing would be misplaced. In this specific context, the individuals don’t have the capacity to identify with group. If some of the members tried to do their part in the morally commended collective action, they may deserve praise for that, and other individuals may deserve blame for not doing so, but it does not seem right to blame the collective. This could be what happens in some real life cases where bystanders witnessing an assault or an accident fail to get together and help victims. Our account will then be more restrictive with assigning collective moral obligations in such types of cases than some other accounts. If we go for the latter option, then there would still be a collective obligation (assuming that each has the situation-specific capacity to group identify), but the members fail to act on it because each or all of them have the wrong moral view (in general, moral ignorance isn’t an excuse though). At this point, we aren’t sure which way to go here, but we are grateful for your comments. This is something we need to think more about.
Cheers
Olle & Björn 🙂
Hi Olle and Björn, I’m always frustrated by two features of most discussions of collective responsibility that I am familiar with: 1) the (more or less exclusive) focus on ‘structured’ groups with decision procedures; 2) the neglect of epistemic aspects of group constitution – before we can have an obligation to do something about something, we need to be jointly aware of it. Addressing these two in one example, it seems to me that joint attention is or at least can be sufficient to constitute a group subject that can also be a subject of joint obligation. Suppose we see the child is about to fall from the table and exchange a ‘sharing look’. We may have never met before, but this is enough to establish us as a plural subject that should act by jointly intervening somehow. Reasoning need not come into it. Or one might try to draw individuals’ attention to something so as to make them jointly aware of it. Do you agree with this? Some things you said pointed in that direction, but I wasn’t entirely sure. And your example remained somewhat hard to pin down for me because you stipulated that the subjects have common knowledge of the situation, but did not give any indication how. So I think it might be helpful if you could clarify what role joint attention plays on your view.
Thanks Michael! You may be right in the sense that joint attention in some situations could trigger group identification, but we doubt that joint attention would always be sufficient. It seems to us that Ale and Olivier could have common knowledge of the alternatives because they jointly attend to the features of the situation, but still lack the capacity to group identify and therefore also lack the obligation. But maybe you think that joint attention in itself involves a collective perspective? That is worth considering, but we take it that many accounts of joint attention understand the phenomenon in terms of mutual attention rather than in terms of a basic we-perspective.
Regarding your remark that “[r]easoning need not come into it”: The boring answer is that reasoning is no more nor no less important for collective obligation than it is for individual obligation. Arguably, the crucial feature of team reasoning is the shift in agential perspective rather than the deliberative process itself. So we agree with you that recognizing that we have an obligation to do something might be more a matter of perceiving an affordance for joint action than consciously deliberating to reach a practical conclusion.
Cheers, Olle & Björn
Thanks for the response, Björn and Olle! I have the feeling we do agree, and yes, I think we should have a notion of joint attention that means more than just mutual perception. It’s true that some philosophers attempt to reduce JA to that, but intuitively jointness is more than just mutuality, and the literature in developmental psychology also assumes a stronger notion – e.g. Carpenter, I am curious what she will say about that tomorrow!
Hi Olle and Bjorn – I’m going to miss your discussion unfortunately so I will ask a question here. I had a version of the question you addressed at the end: is this just a defence of *speaking as if* the group has an obligation? In response, you said that on Hare’s account, to have an obligation just is to be addressable as holding the obligation. But it wasn’t clear to me that it’s really the group that we’re addressing with the obligation. It seems rather that the individuals are the ones being addressed, since it is the individuals who must adopt the group’s perspective and engage in we-reasoning. In which case, according to the Hare view you mention at the end, it seems to be the individuals that have the obligation (where the obligation is to adopt the group’s perspective).
(P.S. I really liked how you turned to ‘look’ at the other while the other was talking!)
Thanks for this comment, Steph! On the Harean view that we toy with here, the function of assigning an obligation is to issue a moral prescription, expressing a strong moral preference about what the addressee should do. In this case what we prescribe is not that each individual must group identify, but that the group must perform a certain collective action that no individual can perform. So, the imperative is addressed to the group. (Imperatives can be addressed to several agents collectively: consider a bossy relationship counsellor uttering ”Get married!” to a client couple, or ”Quarrel!” uttered by a theatre director to the actors rehearsing on stage.) The point we try to make is that addressing the yet unstructured group in this way will not make sense unless we assume that the individuals in the group at least has the capacity to identify with the group and view the situation from the group’s perspective.
Suppose that Ale and Olivier each has the capacity to group-identify and team reason in Burning Building, but despite this, neither of them adopts the group perspective in Burning Building. However, each plumps for going to the first room and they thereby accidentally end up saving the two children. Here, Ale and Olivier have acted in accordance with the moral imperative addressable to them (although they haven’t acted for the reasons that make their joint action obligatory). If the imperative was rather a distributive one addressed to each as you suggest (”Each of you, adopt he group perspective and do your part of saving the two children!”), then they wouldn’t have acted in accordance with the imperative, which to us seems to be the wrong result.
PS. Nothing in this answer really depends on Hare’s prescriptivist view, btw.
Cheers
Olle & Björn